Diminishing the risk of nonrelapse mortality in hematopoietic stem cell transplantation: Prediction of exposure to the cyclophosphamide metabolite carboxyethylphosphoramide mustard.

Abstract

Our objectives were (1) to develop a population pharmacokinetic model for cyclophosphamide, 4-hydroxycyclophosphamide, and carboxyethylphosphoramide mustard (a reporter for nonrelapse mortality) in hematopoietic stem cell transplantation patients and (2) to validate a Bayesian approach to dosing. In this study 147 patients received intravenous infusions of 60 mg. kg -1. d -1 cyclophosphamide for 2 days, followed by 12 to 14.4 Gy total body irradiation. A population model was developed to fit concentration-time data of cyclophosphamide and metabolites. Bayesian prediction of the area under the curve (AUC) was validated by dividing the data set into an index set (98 patients) and validation set (49 patients). Parameters from the index data set were used as priors. Cyclophosphamide elimination was best described by noninducible and inducible routes producing 4-hydroxycyclophosphamide. Induction was described by a zero-order maximum fold of induction-type increase in enzyme level. The prediction of the AUC of carboxyethylphosphoramide mustard was clinically accurate and precise (mean prediction error = -3.5% and root mean squared prediction error = 12.2%) with data limited to 5 to 6 points obtained in the first 16 hours. However, the AUC of 4-hydroxycyclophosphamide was overestimated (mean prediction error = 16.9%-23.6%). Several alternative models did not improve the result. The integrated mechanism-based model describes the pharmacokinetics of cyclophosphamide and carboxyethylphosphoramide mustard. Accurate modeling of 4-hydroxycyclophosphamide is limited by its chemical instability. Exposure to carboxyethylphosphoramide mustard could be accurately and precisely predicted with minimal data obtained over a 16-hour period after the first dose, offering the potential of pharmacokinetically guided dosing to reduce the nonrelapse mortality rate.